| |||||||||||||
On DesignersOn DesignersPosted Oct 1, 2009 23:23 UTC (Thu) by jmm82 (guest, #59425)In reply to: On Designers by zlynx Parent article: LinuxCon: Some advice from Uncle Dirk
Sadly, A new paint job sells a house quicker than fixing the septic tank, but the new owners will smell the backed up S!@@!, eventually. :}
Many Linux projects are in tender need of some love and care, but look at Network Manager. The program is clearly much more user friendly to most *normal* people than ifup/ifdown, still it gets A LOT of resentment for breaking the elegant, simplistic world we can UNIX. Granted it hijacks your routing tables and causes some networking bugs to be so abstracted that only a few people can solve them, but when it works it really is nice.
Simplicity for users comes at the expense of programming complexity. Which leads to the big question. Do all Linux programmers really want the rest of the world to like their masterpiece or will the masterpiece be destroyed in the process leaving a system which resembles the Dredded Micro *cough* *cough* thingie which we left so long ago in search of tranquility.
(Log in to post comments)
On Designers Posted Oct 7, 2009 8:12 UTC (Wed) by alext (guest, #7589) [Link]
But that is thinking of the designer as a decorator rather than the person
who over sees the sum of all the parts and designs a system and by decomposition the parts.
Design is making components that work together by design rather than hacked
On Designers Posted Oct 7, 2009 14:55 UTC (Wed) by jmm82 (guest, #59425) [Link]
Yes, true. It is nearly impossible to find a designer which can design _both_ the internal systems workings and the *eye candy*. A strong large project requires a minimum of three good designers.
1. Systems integration designer
Also, the GUI should be easy to replace with out changing the low-level code. Most projects have #2, but few have all three.
Network Manager is not be a good example because a. it is trying to tackle a problem that is close to impossible and that is to cover the needs of all network use cases. When someone comes across a use case that is not supported they say integrate it with NM or too bad. While I support NM as a project necessary to advance Linux, I disagree with this philosophy. b. Most issues I hear about are due to a lack of documentation and issues with drivers.
|
|||||||||||||